2 .\" $FreeBSD: src/share/man/man4/polling.4,v 1.1.2.4 2003/04/14 08:58:02 maxim Exp $
3 .\" $DragonFly: src/share/man/man4/polling.4,v 1.7 2007/02/24 08:27:38 swildner Exp $
10 .Nd device polling support
12 .Cd "options DEVICE_POLLING"
14 Typically, devices generate interrupts when they need attention
15 from the CPU. Device polling
17 for brevity,) refers to a technique for handling devices that does not
18 rely on interrupts. Rather, it lets the CPU poll devices periodically
19 to service their needs.
20 This might seem inefficient and counterintuitive, but when done
23 gives more control to the operating system on
24 when and how to handle devices, with a number of advantages in terms
25 of system responsivity and performance.
29 reduces the overhead for context
30 switches which is incurred when servicing interrupts, and
31 gives more control on the scheduling of the CPU between various
32 tasks (user processes, software interrupts, device handling)
33 which ultimately reduces the chances of livelock in the system.
34 .Sh PRINCIPLES OF OPERATION
35 In the normal, interrupt-based mode, devices generate an interrupt
36 whenever they need attention.
38 context switch and the execution of an interrupt handler
39 which performs whatever processing is needed by the device.
40 The duration of the interrupt handler is potentially unbounded
41 unless the device driver has been programmed with real-time
42 concerns in mind (which is generally not the case for
45 Furthermore, under heavy traffic, the system might be
46 persistently processing interrupts without being able to
47 complete other work, either in the kernel or in userland.
50 disables interrupts by polling devices at appropriate
51 times, i.e., on clock interrupts, system calls and within the idle loop.
52 This way, the context switch overhead is removed.
54 the operating system can control accurately how much work to spend
55 in handling device events, and thus prevent livelock by reserving
56 some amount of CPU to other tasks.
62 .Va kern.polling.enable
63 whereas the polling frequency is controlled by the
66 .Va kern.polling.pollhz
67 whose range is 1 to 30000 (2000 is the default value).
68 The percentage of CPU cycles reserved to userland processes is
72 .Va kern.polling.user_frac
73 whose range is 0 to 100 (50 is the default value).
77 is enabled, and provided that there is work to do,
79 .Va kern.polling.user_frac
80 percent of the CPU cycles is reserved to userland tasks, the
81 remaining fraction being available for device processing.
85 also changes the way network software interrupts
86 are scheduled, so there is never the risk of livelock because
87 packets are not processed to completion.
89 There are other variables which control or monitor the behaviour
90 of devices operating in polling mode, but they are unlikely to
91 require modifications, and are documented in the source file
92 .Pa sys/kern/kern_poll.c .
95 requires explicit modifications to the device drivers.
96 As of this writing, the
113 devices are supported, with other in the works.
114 The modifications are rather straightforward, consisting in
115 the extraction of the inner part of the interrupt service routine
116 and writing a callback function,
119 to probe the device for events and process them.
121 conditionally compiled sections of the devices mentioned above
124 In order to reduce the latency in processing packets,
125 it is advisable to set the
128 .Va kern.polling.pollhz
131 Device polling was introduced in February 2002 by
132 .An Luigi Rizzo Aq luigi@iet.unipi.it .